ork.AI
Bold. Scary. Uncontrollable. These are often the terms used when people speak of using artificial intelligence (AI) in banking operations. Yet, AI is not new in banking. Compliance programs have been leveraging AI – especially machine learning (ML) and natural language (NL) technologies – with great success since 2016. This is why regulators, as well as compliance leaders at financial institutions, are both promoting and planning for further widespread adoption of AI-based technologies in 2024 and beyond.
The UK’s Financial Conduct Authority (FCA) has adopted AI and expects financial services firms to do the same. According to the FCA’s April 2024 AI Update report, “Our own use of AI is not just changing markets but also the way we regulate. Advanced models can help identify fraud and bad actors, for example.”
In discussing already-proven technologies, the FCA report goes on to state, “We are currently exploring potential further use cases involving Natural Language Processing to aid triage decisions, assessing AI to generate synthetic data or using LLMs to analyse and summarise text.”
In the US, two years have already passed since the Department of the Treasury’s Office of Foreign Assets Control (OFAC) published Sanctions Compliance Guidance for Instant Payment Systems. In its guidance, OFAC encourages the development and deployment of innovative sanctions compliance approaches and technologies to address risks. OFAC stated in the report, “Technology solutions for sanctions compliance, which have advanced significantly in recent years and become more scalable and accessible, can be leveraged to help mitigate a financial institution’s sanctions risk, including with respect to instant payment systems.”
Large banks’ long-term use of AI in compliance operations
No doubt, many of you reading this post will be shocked to learn just how many different subsets of AI have already been proven valuable in AML and sanctions compliance operations. Depending on the use case (and there are many), here are the commonly used subsets of AI used by FIs:
- Machine learning (ML))
- Natural language understanding
- Natural language processing (NLP)
- Natural language generation
- Classification
- Information extraction
- Named entity recognition (NER)
- Question answering
- Summarization
And in 2024, we already have FI customers who have implemented our AML and sanctions compliance solutions that operate using the latest in AI – large language models (LLMs) and Generative AI (GenAI) to improve alert decisioning and alert disposition narratives.
Where and why our Financial Services customers have most used AI
Overwhelmingly, WorkFusion’s FI customers – of which we count 4 of the top 5 US banks – leverage our purpose-built AI solutions in the areas of adverse media screening for KYC, names sanctions screening, transactions screening, and transaction monitoring. They use our AI solutions in these areas because they are where the needs and ROI are typically greatest. It’s easy to see why. Here are three examples:
Adverse media & names Sanctions Screening: The AI-based Digital Worker named Evelyn automates adverse news and sanctions alert review, streamlining time-consuming, error-prone tasks. Specifically, Evelyn dispositions adverse media and customer sanctions alerts as part of KYC or periodic refresh process, intelligently and automatically dispositioning 70 to 80 percent of alerts. Evelyn escalates alerts that require deeper investigation and provides the supporting narrative and documentation for full explainability.
Transaction Screening: The AI-based Digital worker named Tara is purpose-built to automatically dispositions alerts related to the processing of payments, ensuring that payments to and from sanctioned organizations and individuals do not slip through unnoticed. Tara dispositions false positive alerts at a rate of 70 to 80 percent of all alerts and escalates those that require deeper investigation, providing the supporting narrative and documentation for full explainability.
Transaction Monitoring Investigation: The AI -based Digital Worker named Isaac is purpose-built to automate transaction monitoring level one alert review by investigating and evaluating unusual transactions generated from surveillance monitoring systems. Isaac clears away non-suspicious alerts, escalates alerts that require deeper investigation and provides the supporting narrative and documentation for full explainability.
In financial crime compliance, AI normalizes the alert review process – regardless of alert volume surges. So, by leveraging AI, teams reduce manual and repetitive work and experience better quality than they obtain from analysts working overtime or from less qualified, outsourced analysts. AI also frees up analysts to focus on higher-value investigative work and risk analysis.
Pre-built AI solutions that perform specific AML and sanctions job roles give any FI the ability to “hire” AI that works 24/7/365. They quickly resolve staffing challenges and rapidly scale to increase capacity, allowing the compliance team to focus on higher impact investigations and risk analysis.
FIs that can show regulators that they are addressing compliance shortfalls with proven AI results enjoy greater regulatory approval and tend to incur lower penalties and fines due to their demonstrated effort to improve compliance in line with the sophistication of nefarious actors’ use of advanced technologies.
Custom AI solutions deployed by FIs
Going back to 2016, prior to the creation of our AI Digital Workers, banks were deploying WorkFusion AI, most often the specific subsets of ML and NLP (with OCR). Here are two examples:
Instrument Setup: Optical character recognition (OCR) is used to read term sheets and to identify the necessary field values within each term sheet. It commonly extracts 40 or more fields to set up an associated instrument. The solution automatically transforms term sheet content into computer readable text, then applies cognitive intelligence and ML to identify, classify and extract the necessary information.
Employee Archiving: A custom use of AI for the Human Resources department at a large bank leverages OCR to read documents, categorize them and transform them into machine-readable digital documents. From there, integration with the bank’s HR system enables automatic identification of each relevant employee so that any missing details can be added to a document. Human-in-the loop workflow – combined with NLP – enables the solution to call upon HR analysts to quality check newly processed documents. If exceptions exist, they are noted, and the ML component learns this new information for continuous improvement.
As the above examples illustrate, whether referring to custom AI solutions or purpose-built ones for AML and sanctions compliance, AI has been delivering nearly a decade of value to bank compliance operations. Now that the regulators who evaluate your financial institution’s compliance measures support and encourage the use of AI, now is the best time to evaluate how AI can help you scale compliance operations and meet the challenges presented by an ever-more-sophisticated landscape of financial criminals.
To learn more about AI Digital Workers and their use in Financial Services firms today, visit www.workfusion.com.